Singular Value Decomposition and Discrete Wavelet Transform-Based Fingerprint Gender Classification
Nowadays gender classification has immense value in technology and science because it helps to analyze the data easily. Due to uniqueness and reliability of fingerprint they can be used in civilian, industrial, commercial, unique Id of nation (AADHAR card
- PDF / 126,439 Bytes
- 7 Pages / 439.37 x 666.142 pts Page_size
- 3 Downloads / 236 Views
Abstract Nowadays gender classification has immense value in technology and science because it helps to analyze the data easily. Due to uniqueness and reliability of fingerprint they can be used in civilian, industrial, commercial, unique Id of nation (AADHAR card) and even in judicial matters also. There are various methods of fingerprint based gender classification with their own limitations and strengths. In this paper, we presented Singular Value Decomposition (SVD) with Discrete Wavelet Transform (DWT) method for gender classification. SVD has good information packing characteristics and potential strengths in showing the results and DWT has good efficiency and less complexity for the gender identification using fingerprint. Fingerprint images from known gender are obtained for this study. After enhancement, features in frequency domain and spatial domain are obtained by DWT and SVD respectively. Euclidean distance is used to compare with the database. Our proposed method has success rate of around 92 %.
⋅
⋅
Keywords Gender Fingerprint Discrete wavelet transform decomposition Euclidean distance
⋅
⋅
Singular value
1 Introduction A fingerprint forms during pregnancy and remains for whole span of life with the person. If the finger of a person is injured, burned, the prints damages for little period of time. After getting well, it reconstructs as usual. It means there are no changes in the fingerprints. There is no chance to have same fingerprints of two persons even in twins also. Due to uniqueness of fingerprint it can be accepted in G.B. Dongre (✉) CSMSS College of Polytechnic, Aurangabad, Maharashtra, India e-mail: [email protected] S.M. Jagade College of Engineering, Osmanabad, Maharashtra, India e-mail: [email protected] © Springer Science+Business Media Singapore 2017 S.C. Satapathy et al. (eds.), Proceedings of the International Conference on Data Engineering and Communication Technology, Advances in Intelligent Systems and Computing 468, DOI 10.1007/978-981-10-1675-2_1
1
2
G.B. Dongre and S.M. Jagade
civilian, industrial, commercial and unique Id of nation and even in judicial matters also. Fingerprints have immense importance, with the help of which we can easily find out the various properties like age, gender etc. and which will be very much useful for the various applications. Different biometric characteristics-based techniques has been used to differentiate the gender. In this paper, we used SVD and DWT to classify the gender from fingerprint. The 2D-Discrete Wavelet Transform is used to find out the frequency domain vector and Singular Value Decomposition (SVD) is used to find the spatial feature of the non-zero singular values.
2 Literature Review For any fingerprint-based gender determination study, the basic step is fingerprint recognition in which feature extraction is a key point. Bana et al. in [1] proposed image segmentation method for recognition, in which specific area, i.e. RIO of the fingerprint image is obtained using block-based and morphological methods and are operated on. Park et al
Data Loading...
